Living Screener MCP
Review-only MCP server for stock, options, and crypto screening, trade planning, risk checks, backtesting, and journaling, designed to work alongside Robinhood's Trading MCP without placing orders.
README
Living Screener MCP
Review-only hosted MCP for market scanning, trade planning, risk checks, journaling, and postmortems.
This app does not store Robinhood credentials, does not call Robinhood APIs, and cannot place brokerage orders. Robinhood Trading MCP stays separate for broker data, order review, and any future exact user-approved execution.
Current Build
2026.06.08-backtest-rvol
Render Environment
Keep your Finnhub key only in Render:
FINNHUB_API_KEY=<your key>
The checked-in render.yaml defaults to:
MARKET_DATA_PROVIDER=finnhub
MARKET_DATA_INTERVAL=5m
MARKET_DATA_PERIOD=5d
MAX_DATA_STALENESS_MINUTES=1440
REGULAR_MARKET_MAX_STALENESS_MINUTES=15
REVIEW_ONLY=true
PLACE_ORDERS=false
HOSTED_MODE=false
PENDING_BUY_RECHECK_SECONDS=60
SCALP_CANDIDATE_SCORE_THRESHOLD=58
This blueprint uses Render's free web-service plan and does not attach a persistent disk. That keeps the first hosted version inexpensive, but local SQLite journal entries can disappear after Render sleeps, restarts, or redeploys the service. The scanner and safety checks still work normally.
Finnhub is used as the primary quote source. Finnhub's stock candle endpoint may require paid access, so this build safely falls back to yfinance for candles when Finnhub returns no candle data. Market-closed freshness handling accepts the latest regular-session close during overnight and weekend periods instead of rejecting it just because the market is closed.
Options trade plans are gated by separate options-chain validation. The MCP checks bid/ask spread, volume, open interest, expiration risk, and max loss before any review-only options plan can be produced. If those fields are missing or weak, the result is NO_TRADE_PLAN.
Use review_candidate_for_options when a stock candidate appears. It runs the stock setup review and the options-chain gate together so the result is either NO_TRADE_PLAN or REVIEW_ONLY_OPTIONS_READY.
Pending buy orders should be reconsidered after 60 seconds. Use review_pending_buy_order with the ticker, submitted timestamp, and limit price. It cannot cancel or replace anything; it returns whether the pending buy is still valid for review or needs reconsideration.
Use run_scalp_scan for a broader review-only moving-equities scan. It uses a larger liquid/options-friendly universe and scores absolute movement, recent momentum, relative volume, and VWAP alignment. This mode is wider, but it still only produces review candidates; options and pending-order gates remain separate.
Scalp options reviews have a hard relative-volume gate. If a scalp candidate's relative volume is below SCALP_MIN_RELATIVE_VOLUME, review_candidate_for_options returns WATCH_ONLY instead of REVIEW_ONLY_OPTIONS_READY, even when the option chain is liquid.
Backtests now apply the scalp relative-volume gate when the engine name includes scalp, and return aggregate horizon summaries for 15m, 30m, 1h, and the configured close/forward window.
Verify Live Deploy
Open:
https://living-screener-mcp.onrender.com/version
https://living-screener-mcp.onrender.com/tools
https://living-screener-mcp.onrender.com/config
https://living-screener-mcp.onrender.com/health
/version should show:
{
"build_version": "2026.06.08-backtest-rvol",
"market_data_provider": "finnhub",
"has_finnhub_api_key": true,
"can_place_order_from_this_mcp": false
}
Local Run
python -m venv .venv
.\.venv\Scripts\python.exe -m pip install -r requirements.txt
.\.venv\Scripts\python.exe -m unittest discover -s tests
.\.venv\Scripts\python.exe -m uvicorn app.main:app --host 127.0.0.1 --port 8000
MCP Endpoint
https://living-screener-mcp.onrender.com/mcp
Use no auth while HOSTED_MODE=false. Do not send private account data while no auth is enabled.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.